A growing set of policies involve transfers conditioned upon socially desired actions, such as attending school or conserving forest. However, given a desire to maximize the impact of limited funds by avoiding transfers that do not change behavior, typically some potential recipients are excluded on the basis of their characteristics, their actions or at random. This paper uses a laboratory experiment to study the behavior of individuals excluded on different bases from a new incentive that encourages real monetary donations to a public environmental conservation program. We show that the donations from the individuals who were excluded based on prior high contributions fell significantly. Yet the rationale used for exclusion mattered, in that none of the other selection criteria used as the basis for exclusion resulted in negative effects on contributions.